health policy and the disease problem.mapping european healthcare systems

http://esp.sagepub.comJournal of European Social Policy

DOI: 10.1177/0958928709344247 2009; 19; 432 Journal of European Social Policy

Claus Wendt provision and access to healthcare

Mapping European healthcare systems: a comparative analysis of financing, service

http://esp.sagepub.com/cgi/content/abstract/19/5/432 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

can be found at:Journal of European Social Policy Additional services and information for

http://esp.sagepub.com/cgi/alerts Email Alerts:

http://esp.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.co.uk/journalsPermissions.navPermissions:

http://esp.sagepub.com/cgi/content/refs/19/5/432 Citations

at University of Bath on March 15, 2010 http://esp.sagepub.comDownloaded from

Article

Mapping European healthcare systems: a comparative analysis offinancing, service provision and access to healthcare

Claus Wendt*,

University of Mannheim, Germany,and Harvard University, Cambridge, MA, USA

Summary Healthcare systems have been institutionalized to provide healthcare for those in need.Therefore, comparisons should focus in particular on differences in healthcare provision and on howaccess to healthcare services is regulated. This article presents a typology of healthcare systems whichsimultaneously takes into account data on expenditures, financing, provision and access to health-care in 15 European countries. On this basis, three types of healthcare system have been constructedusing statistical cluster analysis: a health service provision-oriented type that is characterized by ahigh number of service providers and free access for patients to medical doctors; a universal cover-age controlled access typewhere healthcare provision has the status of a social citizenship right andequal access to healthcare is of higher importance than free access and freedom of choice; and a lowbudget restricted access type where financial resources for healthcare are limited and patientsaccess to healthcare is restricted by high private out-of-pocket payments and the regulation thatpatients have to sign up on a general practitioners list for a longer period of time.

Key words access to healthcare, cluster analysis, healthcare systems, health policy, typology

Introduction

Healthcare systems provide security against majorlife risks: Not often, but sometimes, it is a matter oflife and death. More usually it represents a powerfulmeans of alleviating the anxiety, discomfort, andincapacity that come from sickness and ill health(Freeman andMoran, 2000: 35). When studying howprotection during illness has been institutionalized indifferent countries, healthcare systems are often dis-tinguished according to their main source of funding.While cross-country comparisons of welfare stateshavemade considerable progress from the early 1990sonwards (Esping-Andersen, 1990; Arts and Gelissen,2001; Scruggs and Allan, 2006), in healthcare systemresearch Social Health Insurance (SHI) types are stillmainly contrasted with National Health Service(NHS) schemes, and the latter has partly been

differentiated into early and late developed NHScountries. Alternatively the names of the foundingfathers, Bismarck and Beveridge, are employed whencomparing these types (Kokko et al., 1998; Marmorand Okma, 1998; Hassenteufel and Palier, 2007).This article argues that comparisons that rely

on broad organizational and financial principles arenot sufficient for gaining a better understanding ofhealthcare systems (see also Marmor et al., 2005).Since healthcare systems serve to provide care forthose in need, comparisons first and foremost haveto concentrate on healthcare provision as well as onhow access to health service providers is regulated.Taking into account information on expenditures,financing, healthcare provision and access to healthservices, this article presents a typology of healthcaresystems which builds on but goes beyond previouscomparative analyses.

*Author to whom correspondence should be sent: Claus Wendt, Mannheim Center for European Social Research,University of Mannheim, A5, 6, 68159 Mannheim, Germany. [email: [email protected]]

The Author(s), 2009. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav Journal of European Social Policy,0958-9287; Vol. 19(5): 432445; 344247; DOI:10.1177/0958928709344247 http://esp.sagepub.com


Based on existing typologies, as outlined in thefollowing section, a set of indicators which representmajor characteristics of healthcare systems is intro-duced in the third part. By using these indicators, 15European healthcare systems were classified withstatistical cluster techniques.1 The construction oftypes helps to better explain how healthcare systemsdiffer from each other. Beyond that, it contributesinsights into the interrelation of the main dimensionsof healthcare systems. More specifically it shows howaccess to healthcare is related to levels of expenditure,the publicprivate mix of healthcare funding, and thedensity of service providers. In the Conclusion thehealthcare system types are discussed and comparedwith earlier typologies. Furthermore, examples areprovided for demonstrating that these types are notonly useful for the understanding of healthcare systemsas such but also provide the basis of further studieswhich may, for instance, focus on their effects on satis-faction, utilization, and health outcomes respectively.

Typologies of healthcare systems

The welfare regime debate provides valuable insightsin conceptual terms but cannot be directly applied forthe comparative analysis of healthcare systems dueto its missing focus on social and healthcare services(Alber, 1995; Bambra, 2005; Wendt et al., 2009). Inorder to close this analytical gap, Bambra (2005) hasintroduced a health decommodification index. Herconcept, however, does not directly cover access tohealthcare providers and can therefore not be takenas a starting point for the typology to be developedin this article.This also holds true for comparative studies on

healthcare systems that focus on modes of governancein order to better understand institutional differencesacross countries. Tuohy (2003), for instance, differen-tiates between agency, contract and networks asmodes of governance in the healthcare arena, whileGiaimo and Manow (1999) draw a distinctionbetween state-led, corporatist-governed andmarket-driven healthcare systems. These and otherconcepts (see Marmor and Okma, 1998; Rothganget al., 2005), however, do not directly link the modesof governance to quantitative data on levels andstructures of healthcare financing and service provi-sion or to institutional data on access to healthcare.Other typologies have a stronger focus on pro-

vision. Field (1973) distinguished healthcare systems

according to the ownership of healthcare servicesand doctors autonomy. An OECD study categorizedhealthcare systems according to the dimensionscoverage, funding and ownership (OECD, 1987)and Frenk and Donabedian (1987) suggested atypology of state intervention in medical carethat is based on the form of state control over theproduction of medical care and the basis for eligi-bility of the population.A conceptual framework that systematically

combines the dimensions of funding, service provi-sion and governance has been introduced by Moran(1999; 2000). His concept of the healthcare stateconsists of the three governing arenas: consumption,provision and production. By referring to Moranstypology, Wendt et al. (2009) suggest combining thedimensions of financing, service provision andregulation with the level of involvement by thestate, non-governmental actors and the market.The result is a taxonomy of 27 healthcare systems,of which three can be identified as ideal types. Thistypology serves to identify differences across coun-tries and changes over time regarding the role of thestate in healthcare in relation to the role of societal-based and private actors.Each of the concepts discussed above (see the

more detailed discussion of health system typologiesin Burau and Blank, 2006; Wendt et al., 2009) covershealthcare provision. However, they do not capturethe number of available providers or regulation ofaccess to healthcare but instead focus on organiza-tional principles on the supply side: ownership ofhealth services (Field, 1973; Frenk and Donabedian,1987; OECD, 1987; Moran, 1999; 2000; Wendtet al., 2009) and doctors autonomy (Field, 1973;Moran, 1999; 2000). As far as access is concerned,it refers to health system coverage (Frenk andDonabedian, 1987; OECD, 1987) but not to patientsaccess to providers. Even if the term consumption isused, the focus is not on patients access to caregiversbut onmore general eligibility criteria for coverage. InMorans concept of the healthcare state, institutionsgoverning healthcare consumption control patientseligibility for access to the healthcare system as wellas the mechanisms which decide on the allocation offinancial resources (Moran, 1999; 2000; Burau andBlank, 2006). The conceptual framework of Wendtet al. (2009) captures regulations of patients access toproviders. However, it is not the strength of regula-tion that is of interest here but who is regulating.

Mapping European healthcare systems 433

Journal of European Social Policy 2009 19 (5)


A strong focus on provision of healthcare,however, is essential to a typology of healthcaresystems that may support comparative analyses of thefunctioning of healthcare systems and their effects onpatients satisfaction, healthcare utilization andhealth outcomes. This does not mean that expendi-ture and financing indictors should be ignored. Theprovision of health services requires funding thattoday in some European healthcare systems exceedsone-tenth of the GDP. Furthermore, the mode offinancing is not only an indicator of the role of thestate in healthcare. Private out-of-pocket paymentsalso have an influence on patients access to serviceproviders andmay discriminate against lower-incomegroups. In the following, a set of indicators is dis-cussed that captures major health system characteris-tics. The selection of indicators follows literaturewhich considers financing, health service provisionand regulation as the main dimensions of healthcaresystems (Moran, 1999; 2000; Rothgang et al., 2005;Powell, 2007). On this basis, 15 European healthcaresystems2 are classified by using cluster analysis. Ourgoal is to construct groups of health systems whichcombine in a typical way expenditure, financing,service provision and access regulation indicators.

Indicators for classifyinghealthcare systems

Healthcare expenditure

Although the control of healthcare expenditure isconsidered to be a major problem in all types ofhealthcare systems, some have turned out to bemore successful in stabilizing healthcare costs thanothers (Freeman and Moran, 2000; Rothgang et al.,2005). Total health expenditure (THE) covers thesum of expenditure for activities in the area of pre-ventive, outpatient and inpatient healthcare, caringfor persons with chronic illness etc. as well asadministering the healthcare system (OECD, 2007).THE can either be calculated as a percentage ofGDP, indicating the level of resources a society iswilling to spend on the provision of healthcare, or itcan be calculated in monetary units per head of thepopulation, indicating the amount of money a societyinvests on average in the health of its members. Inthis article, THE is measured in US dollars per headof the population by using purchasing power parities(PPP/general deflator) since the amount of money

actually spent on healthcare, which is related tothe countrys economic position, is certainly moreimportant for the functioning of the healthcaresystem than the relative level of health expenditure.As argued by Wendt and Kohl (2009), however,there is only a weak correlation between the finan-cial resources invested in a nations health and thelevel of health employment, possibly due to differ-ences in prices and health providers incomechances. Therefore, not only monetary input but alsohealth employment indicators (see below) are to betaken into account when constructing healthcaresystem types.

Healthcare financing

In some comparative studies, the mode of financingis taken as the main or even sole indicator forclassifying healthcare systems. Doubtless, it is veryimportant to patients whether they are entitled tohealthcare on the basis of earmarked social insur-ance contributions, private payments or citizenship(which in general means tax financing) (Mossialosand Dixon, 2002). While information on the modeof entitlement will be covered under the dimensioninstitutional characteristics, in the financingdimension the public share of total health funding(in percent) and the share of private out-of-pocketpayments (in percent of total health financing) willbe included. The share of public funding can betaken as an indicator for the interventional powerof the state (Alber, 1988). The hypothesis that thecapacity to stabilize healthcare costs is greater thehigher the share of public funding is supported bythe fact that today there is a strong negative corre-lation between THE in percent of GDP and theshare of public funding (Wendt and Kohl, 2009).Regarding access to healthcare providers, the shareof public funding indicates to what extent it is con-sidered a public responsibility to guarantee entry forthose in need of medical treatment. For the individ-ual patient, a second indicator of the financingdimension is highly relevant (especially concerninghis or her access to healthcare providers): the level ofprivate out-of-pocket payments (measured in percentof THE). Various studies (Rice and Morrison, 1994;Thomson and Mossialos, 2004; Van Doorslaerand Koolman, 2004; Van Doorslaer et al., 2006)have shown how private cost sharing reduceshealth service utilization and increases inequality.

434 Wendt



The higher the share of private out-of-pocket funding,the greater the privatization of risk in the case ofsickness (Hacker, 2004) and therefore, especially forlower-income groups, the barriers to entering thehealth system.

Healthcare provision

Compared to the high attention paid to expenditureand financing, the production side of health services israther neglected in the international health policydebate. This holds despite the fact that healthcare isparticularly labour-intensive and about 70 percent ofthe total healthcare budgets in Western Europe isdirectly related to employment (Dubois et al., 2006b).In recent publications of the EuropeanObservatory onHealth Systems and Policies Series, health employmentin Europe has been given greater attention (Dubois etal., 2006a; 2006b; Rechel et al., 2006). These publica-tions represent what Marmor et al. (2005) label asstapled national case-studies which allow for adetailed description of healthcare providers and there-fore provide the basis for learning about healthdelivery processes in European countries.For cross-national studies and also for the con-

struction of health system types, however, only alimited number of indicators are to be selected thatrepresent the level and structure of health employ-ment in the included countries. The neglect of healthprovision in comparative studies is probably dueto the difficulties of measuring the level of healthservices on the basis of a few pre-selected indicators.Alber (1988), for example, used the density ofmedical doctors and hospital beds as indicatorsfor the quality of healthcare in OECD countries.Compared with these input indicators, the qualityof health service index developed by Kangas (1994)is more complex and takes into account the earningsreplacement ratio of sickness benefits, the coveragerates of healthcare systems, the number of waitingdays, and the length of the contribution periodrequired for the access to benefits. However, whilethis index covers essential social rights elementsof health systems, it does not directly measure theavailability of health services.For a comparison of the level of healthcare provi-

sion, further or, more precisely, different indicatorsare to be included (McPherson, 1990; Freeman, 2000;Figueras et al., 2004). For patients, the availabilityof healthcare providers is crucial and therefore health

employment data should be directly included in healthsystem comparisons. In the current article, four healthemployment indicators have been selected on thebasis of available OECD data. With these data, twohealthcare provider indices have been constructed:one inpatient care index which includes specialistsand hospital nurses and one outpatient care indexwhich includes general practitioners and pharmacists(see also Wendt and Kohl, 2009). The indicesprovide information on whether healthcare systemsrely to a higher extent on primary healthcare (generalpractitioners, pharmacists) or on specialist healthcare(specialists, hospital nurses).

Institutional characteristics

Access of patients to healthcare is not only influencedby private copayments or available service providersbut also by institutional regulations. A preconditionfor receiving health services is that (potential) carereceivers are covered by the health system. However,since European systems include, with few exceptions,the total population, it is not really meaningful touse the coverage rate for classifying health systems(see, however, Bambra [2005], who takes the cover-age rate into account when calculating a healthdecommodification index).As a first institutional indicator with an effect on

patients access to the healthcare system, the mode ofentitlement is considered. Possible bases of entitle-ment are citizenship, social insurance contributions,private insurance contributions or proven need(Mossialos and Dixon, 2002). Compared to provenneed or entitlement on the basis of citizenship,private and social insurance might stimulate a highertake-up rate of health services in return for contribu-tion payments. While the US healthcare system canbe taken as an example of private insurance being themain basis of entitlement, in Europe it is either citi-zenship or social insurance. Even the most inclusivehealthcare systems cover parts of the population onthe basis of proven need or exempt them fromprivate out-of-pocket payments. Such details cannotbe included in this comparative analysis where thefocus will be on the main mode of entitlement.As a second indicator, the remuneration of doctors

is included. Doctors can be reimbursed on the basisof fee-for-service, per case, per capita (the number ofpatients on his or her list), or by a salary. The controlover doctors income is highest when paying a salary




436 Wendt


and lowest under a fee-for-service scheme (Culyer,1990; Groenewegen et al., 2002). The reimbursementmethod also affects how doctors alter their workloadand it seems fairly clear that fee-for-service methodsresult in both more active treatment and higherincomes of doctors (Culyer, 1990: 38). The remu-neration method indicates the degree of doctorsautonomy from state control (Moran, 1999) andpresumably also influences the level of healthservices provided. Whereas a fee-for-service paymentmay set an incentive for the doctor to see his or herpatients as often as possible, a reimbursement percapita or a fixed salary might set an incentive forreducing the workload (Rice and Smith, 2002).As a third institutional indicator, the regulation of

patients access to healthcare providers is included(see also Reibling, forthcoming). This indicator cap-tures whether patients have a free choice of doctorsor whether they have to sign onto the list of a certaingeneral practitioner (GP) for a longer period of time(gatekeeping system or family doctor principle)(Saltman, 1994; Rico et al., 2003). This indicates thedegree to which patients access to GPs is regulated.Furthermore, access to specialists can be restricted.Patients can have free choice and direct access to spe-cialists. Alternatively a referral by a GP can berequired. In a third type of system, people may skipthe referral system to specialist treatment by accept-ing additional copayment (Reibling and Wendt,2008). For constructing health system types, theseindicators are combined to an access regulationindex which ranges from free choice of doctors(no regulation) on the one side to strict gatekeep-ing on the other, with patients having to sign on aGPs list and needing a referral to specialists. Table 1summarizes the indicators used for the analyses.

Data and analysis

Data

In this article, types of healthcare systems were con-structed by using quantitative data (see Table 2) onthe level of total health expenditure, the share ofpublic funding, the level of private out-of-pocketpayment and the level of health employment. For themeasurement of health service provision, two health-care provider indices were calculated. By using factoranalysis (see annotation in Table 2), two indicatorsfor specialist healthcare, one indicator for primary T

able1

Indicatorsforthecomparativeanalysisofhealthcaresystems

Pub

licp

riva

tePatient

sac

cess

Hea

lth

mix

ofhe

alth

Priva

tiza

tion

Hea

lthserv

ice

Ent

itlemen

tto

Rem

uneration

toserv

ice

Dim

ension

expe

nditur

efina

ncing

ofrisk

prov

ision

health

care

ofdo

ctor

spr

oviders

Indicator

THEper

Publicfunding

Privateout-of-

Indicesof

EntitlementonthebasisRemunerationof

Access

(seedatain

headofthe

in%ofTHE

pocketfunding

healthcare

ofcitizenship,social

GPsonthebasis

RegulationIndex

Table2)

populationin

in%ofTHE

providers

insurancecontributions,offee-for-service,

US$(PPP)

privateinsurance

percapita,

contributions,ordirect

costpercase

out-of-pocketpayments

orasalary

Description

Averagelevel

IndicatoroftheShareoffunding

Indicatorsforthe

TheindicatordescribesThemethodof

Itiscovered

ofmonetary

degreeofpublic

thatisdirectly

levelofhealthcare

themainmodeof

remuneration

whetherpatients

inputsinvested

responsibilityto

coveredbythe

providers.Theindicesentitlement.InEuropean

indicateswhether

havefreeaccessto

inapersons

guaranteeaccessindividual;barrier

areconstructed

healthcaresystems

doctorshavean

GPsandwhether

health

tohealthcare

toentryespecially

byusingdataon

patientsare,ingeneral,

incentiveto

accesstospecialists

forlow-income

specialists,nurses,

coveredeitheronthe

provide

requiresareferral,

groups

general

basisofcitizenshipor

high-volume

additionalcopayment

practitioners,

socialinsurance

healthcare

orisfree

andpharmacists

contributions




Table2

Institutionalcharacteristicsofhealthcaresystems,2001

Institut

iona

lind

icator

sHea

lthex

pend

itur

ean

dHea

lthc

arepr

ovider

privatepa

ymen

tindicesd

Accessregu

lation

PHEbin

private

inpa

tien

tEnt

itlemen

tAccess

THEape

r%

ofO

OPcin

%ca

reou

tpatient

toRem

uneration

GP

Accessto

regu

lation

capita,U

S$THE

ofTHE

inde

xca

reinde

xhe

alth

caree

ofGPsf

registra

tion

specialis

tsinde

xg

1.Austria

2,898

75.7

17.0

109.9

120.2

00

skip&pay

12.Belgium

2,452

76.6

22.2

84.6

193.0

00

skip&pay

13.Denmark

2,561

82.7

15.9

105.2

53.7

11

+referral

34.Finland

1,861

75.9

19.7

79.9

136.4

12

+referral

35.France

2,649

78.3

7.5

90.2

163.7

00

free

06.Germany

2,754

79.3

11.5

120.8

102.6

00

free

07.GreatBritain

2,034

83.0

11.0

91.4

73.3

11

+referral

38.Greece

2,178

47.4

42.4

111.6

68.4

12

free

09.Ireland

2,151

73.6

12.5

107.0

80.1

11

+skip&pay

210.Italy

2,188

74.6

22.1

101.0

122.6

11

+referral

311.Luxembourg

3,270

87.9

6.5

120.2

85.5

00

free

012.Netherlands

2,525

62.8

8.7

109.2

42.0

01

+referral

313.Portugal

1,685

71.5

23.2

68.6

74.3

12

+referral

314.Spain

1,617

71.2

23.9

91.2

110.4

12

+referral

315.Sweden

2,409

84.9

15.1

109.1

73.8

12

+skip&pay

2Not

es:

aTHE:totalhealthexpenditure.

bPHE:publichealthexpenditure.

cOOP:out-of-pocketpayments.

dInafirststepallavailableOECDdataonhealthcarepersonnelhavebeenincludedintheanalysis(specialists,nurses,generalpractitioners,dentists,pharmacists).

Theresultofanunrotatedprincipalcomponentfactoranalysiswasthattwofactorsaccountedfor64%ofthevarianceoftheincludedvariables.However,the

uniquenessofdentiststurnedouttobecomparativelyhigh.Therefore,asecondmodelwascalculatedwithoutdentists.Inthismodel,twofactorsaccountedfor

75%ofthevariance.Thefirstfactorcapturesinpatienthealthcarewithanegativecorrelationbetweenspecialistsandnurses.Thesecondfactoraccountsfor

outpatienthealthcarewithapositivecorrelationbetweengeneralpractitionersandpharmacists.Basedonthefactoranalysiswedecidedtoconstructoneinpatient

careindexandoneoutpatientcareindex.

eCodingforentitlement:SocialInsurance=0;Citizenship=1.

fCodingforremuneration:fee-for-service

=0;capitation

=1;salary

=2.

gCodingforindexconstruction:

=0;

+=1;free

=0;skip&pay

=1;referral

=2.

Sour

ces:OECD(2007);ReiblingandWendt(2008).


438 Wendt


healthcare, and one indicator for pharmaceuticalhealthcare have been selected. These indicators wereaggregated into healthcare provider indices in thefollowing way: first, the raw values for the includedindicators, expressed per 1,000 of population, werestandardized and recalculated as percentages of theEU15 average. The respective index was then calcu-lated as the average value of two health providerindicators. All indicators were weighted equally, thusgiving inpatient healthcare (specialists and nurses)and outpatient healthcare (GPs and pharmacists) which can both be considered as unique and indis-pensable parts of the healthcare system the sameimportance. Furthermore, institutional indicatorshave been selected that cover: (a) the mode ofpatients entitlement to healthcare; (b) the methodfor reimbursing doctors; and (c) patients access tohealth service providers (see Table 2). The latter twohave an influence on the doctorpatient relationshipand the treatment of patients (see Kuhlmann, 2006;Stevenson, 2006).

Analysis

On the basis of data and information included inTable 2, cluster analysis was used to identify certaintypes of healthcare systems and to group countries indifferent clusters (for comparing welfare states byusing cluster analysis see Obinger and Wagschal,1998; Kautto, 2002; Powell and Barrientos, 2004;Jensen, 2008). Cluster analysis aims to group cases bysimultaneously taking a number of selected character-istics into account. Usually the analysis groups cases(here, countries) such that it maximizes homogeneitywithin clusters and maximizes heterogeneity betweenclusters. Ideally countries within clusters should bemore similar to each other than to any country ofanother cluster across all their characteristics. In thepresent analysis, agglomerative hierarchical cluster-ing methods were used (see Everitt et al., 2001).They start out with each country forming a cluster ofits own, and then gradually join countries to formclusters of similar countries until finally all casescome together within one group. Once a country hasbeen allocated to a cluster, it remains within this initialcluster. The result presented in Figure 1 is based onaverage linkage cluster analysis with the Gower dissim-ilarity coefficient since a mix of binary and continuousdata was included. In order to check the stability ofcluster solutions, several other procedures were used

(single and complete linkage, ward method and wav-erage linkage; for methodological details see Gower,1986; Everitt et al., 2001). All procedures (exceptcomplete linkage, where Sweden was considered as adeviant case) created three identical clusters, and thedevelopment of the level of homogeneity (as expressedin the distance coefficient or similarity coefficientrespectively) within country groupings suggested thatthree clusters best represent the structure of the data.With the transition to a two-cluster solution there wasa severe increase in heterogeneity, indicating the com-bination of unlike entities.3

As shown in Figure 1, Greece and the Netherlandscannot be grouped in any of the three clusters. TheNetherlands (before 2006) seems to be unique due tothe high share of private funding, a low level of out-patient healthcare, entitlement on the basis of socialinsurance contributions and comparatively strictaccess regulation. Greece is characterized by thehighest out-of-pocket payments but, in contrast toother Southern countries, has little legal regulationof access to healthcare providers. However, asDavaki and Mossialos (2005) argue, the privatehealth sector is of great importance, and servicesobtained in this sector are related with informalpayments and bribes which restrict access to health-care services for certain population groups.The remaining countries can be classified as follows

(see Table 3):

Cluster 1 consists of Austria, Belgium, France,Germany and Luxembourg (which are all socialinsurance countries4). This type can be describedby a high level of total health expenditure andalso a high share of public funding. The share ofprivate out-of-pocket funding is moderate. Thehigh level of health expenditure is translated intoa moderate level of inpatient and a high level ofoutpatient healthcare. Countries of this clusterare also characterized by a high level of auton-omy of self-employed doctors and high freedomof choice for patients.

Cluster 2 covers Denmark, Great Britain, Sweden(which are all early developed NHS countries),Italy (late developed NHS) and Ireland (no fullyinstitutionalized NHS in 2001). This type is char-acterized by a medium level of total health expen-diture. The share of public health funding is high,and private out-of-pocket funding is moderate.Compared to Cluster 1 the level of inpatient




healthcare providers is similar but the outpatientprovider level is particularly low. The access todoctors is highly regulated, and doctors face strictregulation regarding their income chances.

Cluster 3 includes Portugal, Spain (which arelate developed NHS countries) and Finland (witha NHS introduced in the 1960s). This type ischaracterized by a particularly low level of totalhealth expenditure (per capita) which is (exceptfor Finland) related to the weaker economic posi-tion of these countries. Private out-of-pocketpayments are on a high level and institutionalindicators show a high control of patients accessto medical doctors. The inpatient index is lowand the outpatient index is at a moderate level.Since GPs receive in general a fixed salary,income chances are even more highly restrictedthan in Cluster 2.

These types of healthcare system corroborateearlier comparative studies in this field but alsospecify this debate in four respects. First, it is mis-leading to believe that all countries have to begrouped under a certain type. Second, there are twocases which are grouped in different clusters than

would be expected on the basis of other typologies(see discussion below); Italy joins the group ofestablished NHS countries while Finland, due to itslow level of expenditure, high private out-of-pocketfunding, salary payment and strict access regulation,is much closer to Southern European countries thanto its Scandinavian neighbours. Third, the resultsprovide a more detailed description of healthcaresystem types than earlier typologies. They show, forinstance, that in contrast to Cluster 3, the estab-lished NHS countries of Cluster 2 are regulatingaccess to healthcare in a way which mainly affectsthe level of healthcare provided by GPs and phar-macists but not inpatient healthcare from specialistsand nurses. Fourth, the concept does not implyfrozen types of healthcare system. Changes overtime or the inclusion of further countries will notonly lead to a regrouping of countries but will alsochange the characteristics of a certain type.

Conclusion and discussion

Whenmapping European healthcare systems by usingvarious quantitative and institutional indicators, theidentified clusters partly mirror prior differentiation

Figure 1 Hierarchical cluster analysis: dendrogram using average linkageNote: AT: Austria; DE: Germany; LU: Luxembourg; BE: Belgium; FR: France; DK: Denmark; GB: Great Britain; IE:Ireland; IT: Italy; SE: Sweden; FI: Finland; PT: Portugal; ES: Spain; NL: Netherlands; GR: Greece.


440 Wendt


of healthcare systems into NHS-type countries(with the subgroups of early and late developed NHSsystems) on the one side and SHI-type countries onthe other. In general, the analysis therefore supportsearlier findings and established models of contrastinghealthcare systems.However, earlier typologies (Table 4; see also

Wendt et al., 2009) have been based on eitherfinancing indicators (tax financing versus socialinsurance financing) or organizational tasks (statehierarchy versus self-regulation by corporate actors).By using the dimensions coverage, funding andownership, the OECD (1987) study, for instance,distinguished a national health service model, asocial insurance model and a private insurancemodel. The OECD typology therefore does not offera concept for separating NHS types with differentlevels of health resources, service providers andaccess regulations, respectively. Furthermore, theOECD concept would label the Netherlands as asocial insurance model while the analysis providedin this article demonstrates how difficult it is toclassify a system with social insurance characteristicson the funding side and with comparatively strictregulations in provision and access to healthcare.The typology suggested by Wendt et al. (2009)

focuses on the changing role of the state along thedimensions financing, service provision andregulation. This concept is designed to detect shiftsfrom healthcare offered by public to privateproviders (for-profit or non-profit) and respectivechanges in financing and regulation. Yet differencesacross countries regarding the levels of funding, pro-vision and access to medical care cannot be analysedon the basis of this analytical framework.The typology that comes closest to the solution

offered in this article has been developed by Moran(1999; 2000). His four families of healthcare states,which are based on qualitative judgement, are:entrenched command and control state, thesupply state, the corporatist state and the inse-cure command and control state. The entrenchedcommand and control state is compatible withCluster 2 and it can be argued that the high level ofstate control of doctors autonomy has been used tostabilize healthcare costs, restricting the level of out-patient employment, controlling doctors incomechances and regulating patients access to providers.These are major characteristics of the countriesincluded in Cluster 2. Insecure command andT

able3

Descriptionofclusters

Priva

teou

t-of

-In

dex

Inde

xEnt

itlemen

tAccess

THE

inUS$

Pub

licfu

nding

pock

etpa

ymen

tinpa

tien

tou

tpatient

toRem

uneration

regu

lation

perca

pita

in%

ofTHE

in%

ofTHE

care

care

health

care

ofGPs

inde

x

Cluster1

AT,BE,DE,HighlevelofTHE

Highshareof

Mediumshare

Medium

High

Contributions

Fee-for-service

Low

regulation

FR,LU

(ataverageUS$

publicfinding

ofout-of-pocketinpatient

outpatient

2,805perhead)

(80%

ofTHE)payment(13%

index(105)index(133)

ofTHE)

Cluster2

DK,GB,

Mediumlevelof

Highshareof

Mediumout-of-Medium

Low

Citizenship

Capitation

Mediumto

IE,IT,SE

THE(US$2,269)

publicfunding

pocketpaymentinpatient

outpatient

(exceptSE:salary)strong

(80%

ofTHE)(15%

ofTHE)

index(103)index(81)

regulation

Cluster3

ES,FI,PT

Low

levelofTHE

MediumpublicHighout-of-

Low

Medium

Citizenship

Salary

Strong

(US$1,721)

funding(73%

pocketpaymentinpatient

outpatient

regulation

ofTHE)

(22%

ofTHE)

index(80)

index(107)




control states, by contrast, have never achievedadministrative capacities that are typical of theScandinavian countries and Great Britain. As aresult, barriers to enter the healthcare system aremuch greater, especially for lower-income groups,which in part mirror characteristics of Cluster 3countries. The corporatist state is dominated bypublic law bodies, particularly ambulatory care bydoctors associations (Moran, 2000). This is in linewith characteristics of Cluster 1 countries, whichshow high levels of health expenditure, high serviceprovider levels and privileged income chances fordoctors. Interestingly, the concept proposed byMoran also seems to face difficulties in classifyingthe Netherlands (Burau and Blank, 2006). In contrastto Moran (1999; 2000), however, we argue thatCluster 1 countries are not primarily characterizedby a dominant position of doctors but by compara-tively smooth access of patients to service providers.This may indicate that social insurance agencies are ofhigher importance within the corporate governance

structure than was proposed by Moran, and thatthese agencies have used their position for improvingpatients access to healthcare.Beyond clarifying and, in the case of Moran, com-

plementing earlier concepts, the typology developedin this study enables a detailed description of thethree types which can be labelled as follows:

Health service provision-oriented type. This typeis mainly characterized by its high level andunquestioned importance of service provisionespecially in the outpatient sector. While todaythe mode of entitlement (social insurance con-tributions) is hardly a means for excludingmembers of the population (with some excep-tions), there are various indicators that this typeprovides comparatively smooth access for patientsto service providers. The number of health serviceproviders is high and patients are confrontedwith only modest out-of-pocket copayments.Furthermore, patients have free access to and

Table 4 Typologies of healthcare systems

Dimensions Types of healthcare systems Classification of countries

OECD (1987) Coverage 1. National health service 1. Great Britain funding 2. Social insurance 2. Germany ownership 3. Private insurance 3. United States

Moran (1999); Consumption 1. Entrenched command- 1. Great Britain, Swedenclassification of provision and-control statecountries: see also production 2. Supply state 2. United StatesBurau and Blank 3. Corporatist state 3. Germany(2006) 4. Insecure command- 4. Greece, Italy, Portugal

and-control state

Wendt et al. Role of the state, societal Taxonomy of 27 health 1. Great Britain, Scandinavian(2009) and market actors in: systems with three ideal types: countries

financing 1. State healthcare system 2. No ideal-type; Germany service provision 2. Societal healthcare system represents a societal-based regulation 3. Private healthcare system mixed type

3. No ideal-type: United Statesrepresents a private-basedmixed type

Typology in Health expenditure 1. Health service provision- 1. Austria, Belgium, France,Mapping European Publicprivate mix of oriented type Germany, LuxembourgHealthcare Systems financing 2. Universal coverage 2. Denmark, Great Britain,

Privatization of risk controlled access type Sweden, Italy, Ireland Healthcare provision 3. Low budget restricted 3. Portugal, Spain, Finland Entitlement to care access type Payment of doctors Patients access toproviders


free choice of medical doctors. Since self-employeddoctors are mainly paid on a fee-for-service basis,they have an incentive for more active treatmentto improve their income chances. However, thereare differences to the supplier-dominated healtheconomy, as Moran (2000) has labelled the UScase, since the autonomy of patients and equityof access seem to be of higher importance thandoctors autonomy.

Universal coverage controlled access type. Thistype of healthcare system is mainly characterizedby its universal coverage. Patients access tohealthcare providers, however, is strictly regulatedby the state. This is indicated not only by the lowlevel of outpatient health service providers, butalso by the restricted access to GPs and specialists.The high share of public funding implies strongstate responsibility for the provision of health-care. In general, patients have to sign up on aGPs list for a longer period of time. The remu-neration method of doctors (capitation) alsodoes not set incentives to increase the level ofhealthcare services. However, equity of access issupported since the total population is covered andprivate out-of-pocket payments are particularlylow. In contrast to Moran (2000), Italy joins thegroup of established NHS countries which can berelated to the much higher level of outpatienthealthcare there than in other Southern Europeancountries. Due to its heterogeneous structure, itis in general difficult to classify the Irish systemas a NHS (Nolan, 2008). Based on the indicatorsincluded in this article, however, Ireland hasconsiderable similarities to Great Britain and theScandinavian countries.

Low budget restricted access type. This typeof healthcare system is characterized by a low levelof total health expenditure (per capita). Patientsaccess to healthcare is restricted by high privateout-of-pocket payments and by the regulation thatpatients have to choose their first-contact doctorfor a longer period of time. Furthermore, theinpatient provider level is particularly low. Directprivate payments represent a burden for patients(particularly for lower-income group) and can neg-atively affect equity of access to healthcare. SinceGPs are mainly remunerated on a salary basis, thedegree of doctors autonomy from state controlcan be considered to be even lower than under theuniversal coverage controlled access type.

This classification of health systems depends tosome extent on the selection of cases so that addingfurther nations such as the United States, Switzerlandor Central and Eastern European countries couldreveal different and presumably more than threesystem types. When analysing different points intime it can also be hypothesized that differenttypes of healthcare systems are identifiable andthat countries might change clusters over time. Theconcept introduced in this article, therefore, does notpresume frozen types of healthcare systems but takesinto account that different types can be constructeddepending on time and space.As the results of the cluster analysis are preliminary,

the robustness of the healthcare system types will betested by further analysis. While the classification ofcountries is certainly helpful, it is not an end in itself.Nevertheless, as welfare regime types help us tobetter understand the relationship between socialexclusion and welfare state types, the constructionof this healthcare system typology contributes tothe analysis of the consequences of disparities inhealthcare provision and access. The primary goalof healthcare systems is providing treatment for thosein need. Therefore, the consequences for healthprovision and access to healthcare should be takeninto account when analysing modes of governance(Moran 1999; 2000) or the changing role of thestate (Rothgang et al., 2005).In recent years, a few comparative studies have

become available that focus on the effect of differentinstitutional structures on health outcomes (Conleyand Springer, 2001; Eikemo et al., 2008; Beckfieldand Krieger, 2009). However, these studies concen-trate either on the effects of welfare regimes or eventhe wider political institutional structure, and ingeneral argue that the variation in health is only toa minor extent related to welfare state characteris-tics (Eikemo et al., 2008). Since welfare statetypologies hardly include health system character-istics, typologies of healthcare systems promise todraw a closer link between institutional structuresand health outcomes. And while Van Doorslaeret al. (2006) find only weak effects of institutionalstructures on inequalities in health service use, amore detailed analysis of access regulations shouldfacilitate an explanation of inequalities in healthcareutilization. Furthermore, it can be argued that patternsof satisfaction with healthcare systems should beless related to general welfare state arrangements

442 Wendt



(Gelissen, 2002) and more with specific healthcareinstitutions. It remains to be seen whether suchstudies benefit from a concept where the numberand characteristics of healthcare system types arenot given but depend on time and the sample ofcountries analysed.When learning from other nations healthcare

systems it is essential to take their main characteris-tics into account and not only isolated reform pro-posals. If easy access to healthcare is the primary goal,in the outpatient sector a high level of service providersbecomes especially necessary, which requires incen-tives such as privileged income chances through fee-for-service. Comparatively high costs are theconsequence. If the idea of equal access is moreprevalent, patients visits by general practitioners andspecialists are more regulated and the number of out-patient care providers is much lower.However, in bothcases direct payments by patients are comparativelylow so that a privatization of risk in the case of sicknesscurrently characterizes only a few European countries.The typology presented in this study suggests that cross-national policy learning in the healthcare arena shouldespecially focus on factors which improve patientsaccess to necessary healthcare services.

Acknowledgements

The research reported here has received financialsupport from the German Research Foundation(DFG). I would also like to thank Harvards Mindade Gunzburg Center for European Studies for thetime and intellectual community provided to mewhen I wrote this article as a John F. KennedyMemorial Fellow. A first version of this article waspresented at the ESPAnet Conference 2007, and Igratefully acknowledge the helpful comments andcriticism by the participants, particularly RichardFreeman and Heinz Rothgang, as well as by NadineReibling, Michaela Pfeifer, Monika Mischke, JustinPowell and two anonymous reviewers.

Notes

1 See Jensen (2008) for a similar approach when analysingfinancial transfers and welfare services. Jensen alsoincludes healthcare in his comparative study but onlyby using public health expenditure.

2 The OECD Health Data set provides quantitative datafor further countries. However, in addition to OECDdata, institutional data have been taken from theDFG research project Attitudes Towards Welfare StateInstitutions (MZES/University of Mannheim) which

collects such indicators for the old EU member statesfor 2001. The study has therefore been restricted to 15European countries.

3 After the hierarchical clustering procedures, the robust-ness of this solution was checked with k-means clustering(Powell and Barrientos, 2004; Jensen, 2008). With thismethod, the number of clusters is set by the researcher,and cases are thus selected and recombined to form theoptimal solution regarding homogeneity within clustersin the a priori set number of clusters. Cases can changeclusters during the process of optimizing within-clusterhomogeneity. Thus, the method provides a useful checkfor the results of hierarchical procedures. Again, thethree-cluster solution as reached by these algorithmsshows the highest degree of homogeneity and provesstable using k-means clustering.

4 By referring to social insurance as well to early and latedeveloped NHS systems, similarities to conventionalways of contrasting healthcare systems are indicated.Great Britain and the Scandinavian countries representearly NHS systems while Southern European countrieshave introduced NHS systems only since the late 1970s.

References

Alber, J. (1988) Die Gesundheitssysteme der OECD-Lnderim Vergleich, in M. G. Schmidt (ed.) Staatsttigkeit.International und historisch vergleichende Analysen,pp. 11650. Opladen: Westdeutscher Verlag.

Alber, J. (1995) A Framework for the Comparative Studyof Social Services, Journal of European Social Policy(5) 2: 13149.

Arts, W. and Gelissen, J. (2001) Welfare States, Solidarityand Justice Principles: Does the Type Really Matter?,Acta Sociologica 44 (4): 28399.

Bambra, C. (2005) Cash Versus Services: Worlds ofWelfare and the Decommodification of Cash Benefitsand Health Care Services, Journal of Social Policy34 (2): 195213.

Beckfield, J. and Krieger, N. (2009) Epi + demos + cracy:Linking Political Systems and Priorities to theMagnitude of Health Inequities Evidence, Gaps, and aResearch Agenda, Epidemiologic Reviews, advanceaccess, DOI: 10.1093/epirevmxp002.

Burau, V. and Blank, R. H. (2006) Comparing HealthPolicy: an Assessment of Typologies of Health Systems,Journal of Comparative Policy Analysis 8 (1): 6376.

Conley, D. and Springer, K. (2001) Welfare State and InfantMortality, American Journal of Sociology 107: 768807.

Culyer, A. J. (1990) Cost Containment in Europe, OECDHealth Care Systems in Transition. The Search forEfficiency, pp. 2940. Paris: OECD.

Davaki, K. andMossialos, E. (2005) Plus a Change: HealthSector Reforms in Greece, Journal of Health Politics,Policy and Law 30 (12): 14367.

Dubois, C-A., McKee, M. and Nolte, E. (2006a) HumanResources for Health in Europe, European Observatoryon Health Systems and Policies Series. Maidenhead:Open University Press.

Dubois, C-A.,McKee,M. andNolte, E. (2006b) AnalysingTrends, Opportunities and Challenges, in C-A. Dubois,




M. McKee and E. Nolte (eds) Human Resources forHealth in Europe, European Observatory on HealthSystems and Policies Series, pp. 1540. Maidenhead:Open University Press.

Eikemo, T. A., Bambra, C., Judge, K. and Ringdal, K.(2008) Welfare State Regimes and Differences inSelf-perceived Health in Europe: a Multilevel Analysis,Social Science & Medicine 66: 228195.

Esping-Andersen, G. (1990) The Three Worlds of WelfareCapitalism. Cambridge: Polity Press.

Everitt, B. S., Landau, S. and Leese, M. (2001) ClusterAnalysis. London: Arnold.

Field, M. G. (1973) The Concept of the Health Systemat the Macrosociological Level, Social Science &Medicine 7: 76385.

Figueras, J., Saltman, R. B., Busse, R. and Dubois, H. F. W.(2004) Patterns and Performance in Social HealthInsurance Systems, in R. B. Saltman, R. Busse andJ. Figueras (eds) Social Health Insurance Systems inWestern Europe, pp. 81140. New York: OpenUniversity Press.

Freeman, R. (2000) The Politics of Health in Europe.Manchester: Manchester University Press.

Freeman, R. and Moran, M. (2000) Reforming HealthCare in Europe,West European Politics 23 (2): 3559.

Frenk, J. and Donabedian, A. (1987) State Intervention inMedical Care: Types, Trends and Variables, HealthPolicy and Planning 2 (1): 1731.

Gelissen, J. (2002)Worlds of Welfare, Worlds of Consent?Public Opinion on the Welfare State. Leiden: Brill.

Giaimo, S. and Manow, P. (1999) Adapting the WelfareState the Case of Health Care Reform in Britain,Germany, and the United States, Comparative PoliticalStudies 32 (8): 9671000.

Gower, J. C. (1986) Metric and Euclidean Propertiesof Dissimilarity Coefficients, Journal of Classification3: 548.

Groenewegen, P. P., Dixon, J. and Boerma, W. G. W. (2002)The Regulatory Environment of General Practice: anInternational Perspective, in R. B. Saltman, R. Busse andE. Mossialos (eds) Regulating Entrepreneurial Behaviourin European Health Care Systems, pp. 20014.Buckingham: Open University Press:

Hacker, J. S. (2004) Privatizing Risk Without Privatizingthe Welfare State: the Hidden Politics of Social PolicyRetrenchment in the United States, American PoliticalScience Review 98 (2): 24360.

Hassenteufel, P. and Palier, B. (2007) Towards Neo-Bismarckian Health Care States? Comparing HealthInsurance Reforms in Bismarckian Welfare Systems,Social Policy and Administration 41 (6): 57496.

Jensen, C. (2008) Worlds ofWelfare Services and Transfers,Journal of European Social Policy 18 (2): 15162.

Kangas, O. (1994) The Politics of Social Security: onRegressions, Qualitative Comparisons, and ClusterAnalysis, in T. Janoski and A. M. Hicks (eds) TheComparative Political Economy of the Welfare State,pp. 34664. Cambridge: Cambridge University Press.

Kautto, M. (2002) Investing in Services in West EuropeanWelfare States, Journal of European Social Policy12 (1): 5365.

Kokko, S., Hava, P., Vicente, O. and Leppo, K. (1998)The Role of the State in Health Care Reform, inR. B. Saltman, J. Figueras and C. Sakellarides (eds)Critical Challenges for Health Care Reform in Europe,pp. 289343. Buckingham: Open University Press.

Kuhlmann, E. (2006) Traces of Doubt and Sources ofTrust Health Professions in an Uncertain Society,Current Sociology 54 (4): 60720.

McPherson, K. (1990) International Differences inMedical Care Practices, inOECD Health Care Systemsin Transition. The Search for Efficiency pp. 1728.Paris: OECD.

Marmor, T. R. and Okma, K. G. H. (1998) CautionaryLessons from the West: What (not) to Learn from OtherCountries Experience in the Financing and Delivery ofHealth Care, in P. Flora, P. R. de Jong, J. Le Grand andJ-Y. Kim (eds) The State of Social Welfare, 1997.International Studies on Social Insurance and Retirement,Employment, Family Policy and Health Care, 32750.Aldershot: Ashgate.

Marmor, T. R., Freeman, R. and Okma, K. G. H. (2005)Comparative Perspectives and Policy Learning in theWorld of Health Care, Journal of Comparative PolicyAnalysis 7 (4): 33148.

Moran, M. (1999) Governing the Health Care State. AComparative Study of the United Kingdom, the UnitedStates andGermany. Manchester:Manchester UniversityPress.

Moran, M. (2000) Understanding the Welfare State: theCase of Health Care, British Journal of Politics andInternational Relations 2 (2): 13560.

Mossialos, E. and Dixon, A. (2002) Funding Health Care:an Introduction, in E. Mossialos, A. Dixon, J. Figuerasand J. Kutzin (eds) Funding Health Care: Options forEurope, pp. 130. Buckingham: Open University Press.

Nolan, A. (2008) The Impact of Income on PrivatePatients Access to GP Services in Ireland, Journal ofHealth Services Research & Policy 13: 2226.

Obinger, H. and Wagschal, U. (1998) Drei Welten desWohlfahrtsstaates? Das Stratifizierungskonzept in derclusteranalytischen berprfung, in S. Lessenich andI. Ostner (eds) Welten des Wohlfahrtskapitalismus.Der Sozialstaat in vergleichender Perspektive. Frankfurt a.M./New York: Campus: 10935.

OECD (1987) Financing and Delivering Health Care. AComparative Analysis of OECDCountries. Paris: OECD.

OECD (2007) OECD Health Data 2007. A ComparativeAnalysis of 30 Countries. Paris: OECD.

Powell, M. (2007) The mixed economy of welfare and thesocial division of welfare, in M. Powell Understandingthe mixed economy of welfare. Bristol: Policy Press: 121.

Powell, M. and Barrientos, A. (2004) Welfare Regimesand the Welfare Mix, European Journal of PoliticalResearch 43 (1): 83105.

Rechel, B., Dubois, C.-A. and McKee, M. (2006): TheHealth Care Workforce in Europe. Learning fromExperience. European Observatory on Health Systemsand Policies. Copenhagen: WHO.

Reibling, N. (forthcoming) Healthcare systems in Europe:towards an incorporation of patient access PAnet/JESP

444 Wendt





doctoral researcher prize essay], Journal of EuropeanSocial Policy.

Reibling, N. and Wendt, C. (2008) Access Regulation andUtilization of Healthcare Services, Mannheimer Zentrumfr Europische Sozialforschung Working Paper Series,No. 113. Mannheim: MZES.

Rice, T. and Morrison, K. R. (1994) Patient Cost Sharingfor Medical Services: a Review of the Literature andImplications for Health Care Reform, Medical CareReview 51 (3): 23587.

Rice, N. and Smith, P. C. (2002) Strategic ResourceAllocation and Funding Decisions, in E. Mossialos,A. Dixon, J. Figueras and J. Kutzin (eds) Funding HealthCare: Options for Europe, pp. 25071. Buckingham:Open University Press.

Rico, A., Saltman, R. B. and Boerma, W. G. W. (2003)Organizational Restructuring in European HealthSystems: the Role of Primary Care, Social Policy &Administration 37 (6): 592608.

Rothgang, H., Cacace, M., Grimmeisen, S. and Wendt, C.(2005) The Changing Role of the State in Health CareSystems, European Review 13 (1): 187212.

Saltman, R. B. (1994) Patient Choice and PatientEmpowerment in Northern European Health Systems a Conceptual-Framework, International Journal ofHealth Services 24 (2): 20129.

Scruggs, L. and Allen, J. (2006) Welfare-stateDecommodification in 18 OECD Countries: aReplication and Revision, Journal of European SocialPolicy 16 (1): 5572.

Stevenson, F. (2006) The DoctorPatient Relationship.Interconnections Between Global Processes andInteraction, in C. Wendt and C. Wolf (eds) Soziologieder Gesundheit. Sonderheft der Klner Zeitschrift frSoziologie und Sozialpsychologie 46, pp. 22442.Wiesbaden: VS-Verlag.

Thomson, S. and Mossialos, E. (2004). Funding HealthCare from Private Sources: What are the Implicationsfor Equity, Efficiency, Cost Containment and Choicein Western European Health Systems? Copenhagen:World Health Organization.

Tuohy, C. (2003) Agency, Contract, and Governance:Shifting Shapes of Accountability in the Health CareArena, Journal of Health Politics, Policy and Law28 (23): 195215.

Van Doorslaer, E. and Koolman, X. (2004) Explaining theDifferences in Income-related Health Inequalities acrossEuropean Countries,Health Economics 13 (7): 60928.

Van Doorslaer, E., Masseria, C. and Koolman X. (2006)Inequalities in Access to Medical Care by Income inDeveloped Countries, Canadian Medical AssociationJournal 174 (2): 17783.

Wendt, C. and Kohl, J. (2009) Translating MonetaryInputs into Health Care Services the Influence ofDifferent Modes of Public Policy in a ComparativePerspective, Journal of Comparative Policy Analysis11 (3): 295315.

Wendt, C., Frisina, L. and Rothgang, H. (2009) Health CareSystemTypes. AConceptual Framework forComparison,Social Policy & Administration 43 (1): 7090.


health policy and the disease problem.mapping european healthcare systems

Documents